# Claude Fable on Bedrock — model selection decision matrix

Use this when a stakeholder asks *"should we move everything to Fable 5.1?"* The answer is almost always **no** — route by workload shape, governance constraints, and whether you can accept Anthropic's Covered Model retention.

> Reflects **Claude Fable 5.1** GA on **September 1, 2026** (successor to Fable 5, June 9, 2026). Model IDs, regions, and pricing change — confirm in the [Fable 5.1 model card](https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-anthropic-claude-fable-5-1.html) and [Covered Models policy](https://support.claude.com/en/articles/15425695-covered-models) before committing. Do not quote unpublished list rates.

## 1. Pick the access path first

| You need… | Use | Why |
|-----------|-----|-----|
| Data stays in AWS infra; compose with Guardrails, Knowledge Bases, Agents | **Amazon Bedrock** (`bedrock-runtime` or `bedrock-mantle`) | IAM, PrivateLink, regional residency, existing Bedrock billing tags |
| Anthropic-native experience (projects, Claude Code-style workflows) with AWS billing | **Claude Platform on AWS** | Same models; different console and API surface — see [Claude Platform docs](https://docs.aws.amazon.com/claude-platform/latest/userguide/welcome.html) |
| Unrestricted Mythos-class model (no Fable safeguards) | **Claude Mythos 5.1** (limited access, approved partners) | Not GA for general enterprise — do not architect assuming access |

**We recommend Bedrock** for production apps that already run on AWS governance. Use Claude Platform on AWS for teams that want Anthropic's native UX but need consolidated AWS procurement. Covered Model policies follow the model onto both paths.

## 2. Pick the model tier

| Workload | Lean **Fable 5.1** (current GA) | Lean **Claude Sonnet 5 / Opus 5** | Lean **Nova Pro / Lite** |
|----------|----------------------------------|-------------------------------------|--------------------------|
| Multi-hour coding or research agent (async, self-verifying) | Yes — after `aws_review` sign-off | No — prior models lose coherence on long unbroken runs | No |
| Document-heavy vision (nested tables in PDFs, architecture diagrams) | Yes | Sonnet 5 if docs are short; Opus 5 if reasoning depth + ZDR matter more than vision | No |
| High-volume chat / classification (< 4K context) | No — overkill cost/latency | Haiku-class or Nova Micro | Yes |
| Regulated workload that **cannot** accept 30-day Covered Model retention | **No** | Opus 5 is ZDR by default on Bedrock; Sonnet 5 if retention is acceptable | Yes |
| Prompts touching cybersecurity / bio / chem / health domains | Expect **classifier refusal** or fallback (lower capability) | Sonnet/Opus if you need predictable capability in those domains | N/A |
| Steady production API with strict p99 latency SLO | Cautious — adaptive thinking is always on (default effort: high) | Sonnet 5 for balanced latency | Nova for cost |
| Already pinned on Fable 5 and meeting quality | Bake off 5 vs 5.1 on one lane; do not swap on the version number | N/A | N/A |

## 3. Governance gates (hard stops)

Before routing production traffic to Fable 5.1 (or remaining on Fable 5) on Bedrock, confirm:

- [ ] Legal/security signed off on **Covered Model** retention and Data Retention API mode **`aws_review`** — 30-day minimum input/output retention + safety review. ZDR is **not** the default. (Stale June copy said `provider_data_share` — do not use that string.)
- [ ] You are **not** treating **Enterprise Frontier Safeguards (EFS)** as generally available. EFS rolls out in phases beginning **fall 2026**. A limited-time ZDR transition applies only if Anthropic or AWS offered it to you in writing.
- [ ] Application handles `stop_reason: "refusal"` as a **primary path** — refusal rates are materially higher than prior Claude models. Prompt-stage refusals are not billed; mid-stream refusals bill partial output.
- [ ] Inference profile uses a **region compliant with data residency** — confirm the current matrix on the 5.1 model card (`anthropic.claude-fable-5-1`, `us.anthropic.claude-fable-5-1`, `global.anthropic.claude-fable-5-1`).
- [ ] Cost model uses **live Bedrock pricing** — do not copy unpublished $/MTok from blogs.
- [ ] Long-running jobs have **checkpointing** — Fable can run extended async work; your orchestrator must survive Bedrock timeouts and retries.

## When NOT to adopt Fable 5.1 yet

- Legal has not approved Covered Model retention / `aws_review`.
- You need ZDR today and are not an eligible EFS-transition customer — use Opus 5 on Bedrock instead.
- You have not tested refusal handling on your real prompt distribution — a 3% refusal rate on a 50k/day API is 1,500 user-visible failures.
- The workload is a drop-in replacement for Sonnet 5 on short prompts — you will pay frontier-model prices without using frontier-model duration.
- You were about to design around **Mythos 5.1** — limited access; not a production default.